* Clean up
clear all

/* Set working directory: please set your own
cd "~/Dropbox/JOP third submission/JOP replication/"
*/

* Open dataset
use 01_data/balearic_plus_ciutadella.dta, clear

* 2014/5-2019
reghdfe pp_voteshare post##ep##ciutadella if period > 1 & (ciutadella == 1 | mallorca == 1), absorb(mesa_code_elecspecific) cluster(mesa_code_elecspecific)
est store twoperiods_ddd_mallorca

* Pre vs. post
reghdfe pp_voteshare post##ep##ciutadella if (ciutadella == 1 | mallorca == 1), absorb(mesa_code_elecspecific) cluster(mesa_code_elecspecific)
est store prepost_ddd_mallorca

* Fixed effects for each period
reghdfe pp_voteshare post##ep##ciutadella i.period  if (ciutadella == 1 | mallorca == 1), absorb(mesa_code_elecspecific) cluster(mesa_code_elecspecific)
est store threeperiods_ddd_mallorca

* Make table
esttab twoperiods_ddd_mallorca prepost_ddd_mallorca threeperiods_ddd_mallorca using 03_tables/tabled8.tex, tex se replace mtitles ("2014-2019" "2009-2019" "2009-2019")  keep (1.post 1.post#1.ep 1.post#1.ciutadella 1.post#1.ep#1.ciutadella 2.period) coeflabels (1.post "2019 election" 1.post#1.ep "EP election in 2019" 1.post#1.ciutadella "Ciutadella in 2019" 1.post#1.ep#1.ciutadella "EP 2019 x Ciutadella" 2.period "2014-2015 election") star(* 0.10 ** 0.05 *** 0.01) addnotes("All models include voting station * election type fixed effects" "Standard errors are clustered by voting station * election type") 
